DNA methylation: Methylation status of bovine preimplantation embryos derived from different origins

Although the efficiency of somatic nuclear transfer is low (KATO et al. 1998;

NIEMANN et al. 2002), live offspring has been born after transfer of cloned embryos to recipient animals in eleven mammalian species. An average of 20 to 30% of the bovine reconstructed embryos reach the blastocyst stage and 10 to 15% of the blastocysts transferred to recipients develop to term (COLMAN 2000). In vitro produced embryos are different from their in vivo counterparts in several characteristics including developmental capacity, morphology, metabolism, gene expression patterns and methylation patterns (VAN SOOM et al. 1997a,b; KOO et al.

2002; KANG et al. 2001b; WRENZYCKI et al. 2001, 2002; LAZZARI et al. 2002;

ZHANG et al. 2004; HIENDLEDER et al. 2004). Abnormal fetal phenotypes with a variety of pathologies were summarized as Large Offspring Syndrome (LOS).

According to a current hypothesis, deviations from normal gene expression patterns in preimplantation development are involved in LOS (WRENZYCKI et al. 2001).

Preimplantation embryos may provide an important tool to unravel deregulatory mechanisms related to the Large Offspring Syndrome (LOS).

The present study is the first to analyze methylation patterns of the IGF2 gene in bovine preimplantation embryos. IGF2 was selected because it is known to play a major role during embryonic development. Available studies on bovine embryonic methylation patterns are shown in table 14.

Table 14: Recently published studies on bovine embryonic methylation patterns

Material Method Gene Reference

Bovine tetraploid and diploid blastocysts (NT, PA, IVF)

Bisulfite treatment, PCR, AciI digestion

Satellite I DNA region KANG et al.

(2001a)

Bov-B long interspersed nuclear element sequence

KANG et al.

(2001b) Bovine

preimplantation embryos (NT, IVF)

Immunofluorescence Global methylation pattern DEAN et al.

(2001) Bovine

preimplantation embryos, ICM- and TE-cells (NT, IVF)

Bisulfite sequencing

Epidermal cytokeratin gene Mammary gland-specific

HPLC Global methylation level CEZAR et al.

(2003) Preimplantation

bovine embryos (NT, IVF)

Bisulfite sequencing Bovine neuropeptide galanin gene

KANG et al.

(2003) Bovine liver DNA

from day 80 fetuses (IVF, NT, AI)

Capillary

electrophoresis Global methylation level HIENDLEDER et al. (2004) Bovine day 40 female

fetuses (AI, NT), placenta

Bisulfite sequencing,

RT-PCR IGF2, GTL2, Xist DINDOT et al.

(2004)

Analyses of methylation patterns in bovine preimplantation embryos of different origins revealed that bovine zygotes, collected 19 hours after in vitro fertilization were methylated to 28% and contained only two fully demethylated CpG dinucleotides. At the 4-cell stage the methylation level has decreased to 6% and most of the CpGs analyzed within the IGF2 DMR2 was completely demethylated.

embryos suggests an active demethylation mechanism. Our results support the finding that demethylation of the bovine embryonic genome occurs at an intermediate state (BEAUJEAN et al. 2004) compared to those of mouse and sheep. In the mouse, the male pronucleus is demethylated within a few hours after fertilization and the maternal genome during the first cell divisions (OSWALD et al. 2000; MAYER et al. 2000; HOWLETT and REIK 1991; ROUGIER et al. 1998) whereas in ovine embryos demethylation was observed in the expanded blastocyst for the first time (BEAUJEAN et al. 2004). A putative correlation of demethylation and timing of embryonic genome activation was proposed as an explanation for the time dependent genome-wide demethylation (BEAUJEAN et al. 2004; YOUNG and BEAUJEAN 2004). The embryonic genome is activated in mouse 2-cell embryos (TELFORD et al. 1990) and the parental genomes have to be reprogrammed faster as in bovine embryos in which major embryonic transcription starts at the 8-16-cell stage (KING et al. 1989). It was further proposed that reprogramming of the parental genomes after fertilization is required to attain totipotency (MONK et al. 1987;

HOWLETT and REIK 1991). Mouse blastomeres possess totipotent capabilities until the morula stage when segregation into inner cell mass, endoderm and trophectoderm begins. By the time the murine embryo consists of 32-64 blastomeres, totipotency is lost due to this first differentiation event and the stem cell specific Oct-4 gene is expressed and methylated (ROBERTS et al. 2004; HATTORI et al. 2004). At the murine blastocyst stage, de novo methylation occurs but is restricted to the inner cell mass (DEAN et al. 2001; SANTOS et al. 2002). However, genes responsible for tissue specific differentiation might similarly become methylated in the trophectoderm. The transcription factor Oct-4 plays an important role in the maintenance of pluripotency of the inner cell mass and embryonic stem cells which originate from the ICM. The Oct-4 gene is silent in trophoblast cells due to hypermethylation of the Oct-4 enhancer and promoter region. This region is hypomethylated in embryonic stem cells (HATTORI et al. 2004). Asymmetric methylation patterns were observed in ovine blastocysts due to demethylation in the trophectoderm whereas the ICM remained methylated. This demethylation pattern which is restricted to the trophectoderm was proposed to be critical for differentiation into trophectoderm cells in ovine blastocysts (BEAUJEAN et al. 2004).

Trophectoderm cells are lower methylated than the inner cell mass cells (DEAN et al.

2001) indicating that cell specific methylation levels could have a role in the onset of

tissue specific differentiation. DNMT3A and DNMT3B mRNA levels were low in bovine 8-cell embryos but increased up to the blastocyst stage (WRENZYCKI and NIEMANN 2003). De novo methylation probably begins at the bovine 8-16-cell stage and the de novo methyltransferase is fully active at the blastocyst stage.

Comparison of methylation levels between blastocysts derived from different origins revealed increased methylation levels in blastocysts generated by nuclear transfer from male adult fibroblasts (22%±1.9 SEM). Only androgenetic blastocysts containing a diploid paternal genome showed a higher methylation level (30.6%±2.1 SEM) which is consistent with the finding that the paternal allele is normally methylated at this gene locus. The very low methylation level of parthenogenetic blastocysts (2.3%±1 SEM) is consistent with the finding of low maternal methylation in oocytes (fig. 47) and might represent the normal methylation level of the maternal allele at this gene locus.

Various procedures required to produce embryos in vitro affect embryonic development (NIEMANN and WRENZYCKI 2000; HUMPHERYS et al. 2002; HAN et al. 2003) but did not affect methylation patterns of IGF2 DMR2 in in vitro fertilized (10.1%±0.7 SEM) and blastocysts cloned from female adult fibroblasts (12.4%±1.4 SEM) in the present study. The methylation patterns were similar to those observed in in vivo derived embryos (10.2%±1.2 SEM) and were consistent with results from immunofluorescence studies (DEAN et al. 2001) and gene expression analyses of DNMTs1, 3A and 3B (WRENZYCKI and NIEMANN 2003). The “normal” methylation pattern of female cloned blastocysts could suggest that female embryos are more tolerant to in vitro procedures and/or methylation reprogramming under our laboratory conditions. Female cloned calves from fetal and adult fibroblasts have been born in our laboratory during the past few years (NIEMANN et al. 2002), but no male calf. Presumably, the increased methylation level in male cloned embryos found in this study is the result of incomplete reprogramming. Insufficient demethylation was also reported for male cloned preimplantation embryos at the Bov-B long interspersed nuclear element sequence locus (KANG et al. 2001b). Whether epigenetic failures in cloned embryos occur in a sex specific manner needs further investigations because incomplete reprogramming was also determined in bovine female cloned fetuses (CEZAR et al. 2003). On the other hand, the increased

mRNA expression analyses of the DNMT3A and DNMT3B genes which did not reveal differences in the relative amounts of transcriptional products in 8-cell embryos generated by in vitro fertilization and nuclear transfer (WRENZYCKI and NIEMANN 2003).

Adult fibroblasts employed for the production of cloned blastocysts were highly methylated in both, female and male derived cells (77%±2.2 SEM; 72%±2.9 SEM).

The fibroblasts were serum starved prior to nuclear transfer. No mRNA was detected for the DNMT1 gene in starved somatic cells (BESTOR 2000; WRENZYCKI and NIEMANN 2003). The lower methylation level detected in cloned blastocysts compared to somatic donor cells suggests that the methylation pattern was different in fibroblasts and the donor cell genome became demethylated. Dnmt1 protein was found in cloned murine 8-cell embryos while this protein is normally expressed later in development (CHUNG et al. 2003). Somatic Dnmt1 is absent in oocytes and preimplantation embryos (CARLSON et al. 1992; MERTINEIT et al. 1998; RATNAM et al. 2002). Abnormally high amounts of the somatic form of Dnmt1 were proposed as inhibitors for correct reprogramming in cloned embryos (BESTOR 2000; CHUNG et al. 2003). Incomplete reprogramming of imprinted and non-imprinted genes has often been associated with aberrant embryonic and fetal development of cloned offspring (YOUNG et al. 2001; EGGAN et al. 2001; DEAN et al. 2003; HIENDLEDER et al. 2004). Cloned mouse blastocysts showed aberrant methylation patterns and gene expression in the imprinted genes H19 and Snrpn (Small nuclear ribonucleoprotein N). Even morphologically viable cloned embryos rarely display normal gene expression of imprinted genes (MANN et al. 2003). Bovine somatic cloning is more successful compared to other species such as the mouse. This might be due to an increased tolerance of bovine embryos against disruption of imprinted genes (ZHANG et al. 2004).

In vitro fertilized blastocysts analyzed in the present study possessed similar methylation level compared to the in vivo control blastocysts. HIENDLEDER et al.

(2004) determined the global methylation level of liver DNA in day 80 fetuses derived from nuclear transfer (NT), in vitro fertilization and artificial insemination (AI) and found an increased, albeit non-significant, methylation level in IVF compared to AI control fetuses. The serum supplements in culture media are thought to affect embryonic development at the morphological and transcriptional level (THOMPSON 1997; WRENZYCKI et al. 1999; 2001; BLONDIN et al. 2000; KHOSLA et al. 2001).

The hemi-defined culture medium SOF/BSA which was employed in this study, obviously did not affect reprogramming of the IGF2 DMR2 locus except for male cloned blastocysts. It may represent a suitable culture medium for in vitro production of bovine embryos.

In conclusion, results of this study demonstrate that imprinting is conserved among mammals. Imprinted genes can be bi-allelically expressed from both parental alleles prior to implantation irrespective of differences in placentation as shown in this study for the bovine IGF2, IGF2R and MASH2 genes. This study focused on the analyses of methylation patterns within the bovine IGF2 gene in gametes and preimplantation embryos derived from different origins. Two regions of the 5' untranslated region of the bovine IGF2 gene were sequenced and revealed that sequence identities of the IGF2 gene are conserved among mammals. Most importantly, we identified a Differentially Methylated Region (DMR) in the last exon of the bovine IGF2 gene which presumably corresponds to the DMR2 located in the last exon of the murine Igf2 gene. This DMR is methylated mainly on the paternal allele and becomes reprogrammed after fertilization. The DMR might serve as an important tool to unravel effects of assisted reproduction technologies on embryonic and fetal development. Preimplantation embryos employed in this study were cultured in SOF/BSA medium and their methylation patterns were not perturbated by the in vitro culture system. Male cloned blastocysts had high methylation levels suggesting sex specific failures during methylation reprogramming. The results obtained in this study revealed that in vitro production protocols can affect the methylation pattern of bovine blastocysts. Developmental aberrations in the early embryo might have effects on further development but could be detected by analyses of methylation levels of early preimplantation embryos.

6 SUMMARY

Claudia Gebert

Characterization of gene expression and methylation patterns of the bovine IGF2 gene in gametes and preimplantation embryos of different origins

DNA methylation at cytosine molecules is one of the best-known mechanisms involved in genomic imprinting. Imprinting plays a crucial role during mammalian embryonic development. Insight into the epigenetic phenomenon of imprinting is mainly derived from mice and humans, and information in bovine imprinting is scarce.

The main goal of this study was to analyze the methylation status of the bovine Insulin-like growth factor2 gene (IGF2) in gametes and preimplantation embryos. In addition, expression patterns of the bovine IGF2, Insulin-like growth factor2 receptor (IGF2R) and Mammalian achaete-scute homologue2 (MASH2) genes were characterized in bovine blastocysts.

At first, the expression patterns of day 7 bovine blastocysts and expanded blastocysts were analyzed by semi-quantitative RT-PCR. Embryos were produced by in vitro fertilization (IVF) or parthenogenesis (PA). In vitro culture was performed in the semi-defined medium SOF/BSA (Synthetic Oviduct Fluid/Bovine Serum Albumin). This approach allows the indirect determination of a putative mono- or bi-allelic gene expression by measuring the relative amounts of transcriptional products.

Paternally expressed genes would be expected to show increased relative amounts of mRNAs in in vitro produced embryos (maternal and paternal genome) compared to their parthenogenetic counterparts. In contrast, maternal mRNA expression would be reflected by an increased relative amount of transcriptional products in parthenogenetic (two maternal genomes) compared to IVF embryos. Hatched in vitro produced blastocysts, in vivo collected blastocysts, ICM (inner cell mass) and trophectoderm cells isolated from in vitro fertilized and parthenogenetic blastocysts were included in the analyses. Cell numbers were determined by differential staining for ICM and trophectoderm cells (in vitro fertilized and parthenogenetic blastocysts).

The size of the blastocysts was also measured.

Analyses of the methylation level within the bovine IGF2 gene in gametes and preimplantation embryos required sequencing of the gene due to the scarcity of

genomic sequence information in farm animals. Only the mRNA sequence of the bovine IGF2 gene was available in the GenBank database when this study was initiated. Two fragments within the untranslated region were identified by PCR amplification of bovine tissue DNA using ovine primer pairs. Intron sequences of the translated region were detected by the same approach but with bovine primers. The newly sequenced fragments of the untranslated region and exon 10 were analyzed for their methylation status in isolated DNA from bovine in vitro matured oocytes and frozen/thawed sperm cells. Methylation levels were determined at 27 CG dinucleotides within exon 10 in bovine preimplantation embryos derived from different origins by bisulfite sequencing. Adult fibroblasts, which were used to generate nuclear transfer embryos, were included in the analyses. Bisulfite sequencing is a highly sensitive method to detect methylated and non-methylated cytosine molecules and is based on the ability of bisulfite to convert non-methylated cytosines into thymine whereas methylated cytosines remain unchanged.

The following results were obtained:

1) Differential staining revealed similar cell numbers for in vitro fertilized and parthenogenetic blastocysts (<180 µm) and expanded blastocysts (>180 µm). Increased cell numbers (p≤0.05) in expanded in vitro fertilized blastocysts were attributed to a higher number of trophectoderm cells.

2) Expanded day 7 in vitro fertilized and parthenogenetic blastocysts expressed the bovine IGF2 and IGF2R genes at similar relative amounts.

In contrast, higher relative amounts of transcriptional products of the MASH2 gene were determined for in vitro fertilized blastocysts compared to their parthenogenetic counterparts (p≤0.05).

3) Semi-quantitative gene expression analyses extended to in vitro and in vivo produced blastocysts and isolated ICM and trophectoderm cells revealed increased IGF2 expression in expanded in vitro fertilized blastocysts (p≤0.05). In contrast, the IGF2R gene showed higher relative amounts of transcriptional products in ICM cells isolated from expanded in vitro fertilized blastocysts and in hatched parthenogenetic blastocysts (p≤0.05). Similar expression levels were determined in hatched in vitro fertilized and parthenogenetic blastocysts.

4) Two fragments isolated within the untranslated region of the bovine IGF2 gene showed high homology (>90%) to exons and introns 4 and 5 of the ovine IGF2 gene.

5) CpG islands were detected within the sequences of intron 4 and 5 of the bovine IGF2 gene. In addition, intron 5 contained a promoter region including a TATA box.

6) The DNA fragment of the translated region of the bovine IGF2 gene was homologous to the bovine IGF2 mRNA sequence available in the GenBank database (X53553). Alignment with the already published mRNA sequence identified the non-homologous intron sequences 8 and 9 of the bovine IGF2 gene.

7) Methylation analyses revealed a similar low methylation level of introns 4 and 5 between oocyte and sperm DNA. In contrast, exon 10 was hypermethylated in sperm DNA (99%), but only low methylated in oocyte DNA (16%).

8) DNA methylation within exon 10 was low in bovine in vitro fertilized zygotes and decreased further in in vitro fertilized 4-cell embryos.

9) Day 7 expanded blastocysts collected in vivo, fertilized in vitro or cloned with female adult fibroblasts showed similar methylation patterns. The level of methylation was higher in expanded blastocysts than in 4-cell stage embryos.

10) Bisulfite sequencing revealed an increased and low methylation level within exon 10 in androgenetic and parthenogenetic expanded blastocysts, respectively. Female and male adult fibroblasts were both heavily methylated.

11) Interestingly, male cloned blastocysts were hypermethylated compared to their female counterparts.

The relative amounts of transcriptional products of the IGF2, IGF2R and MASH2 genes detected in in vitro fertilized and parthenogenetic blastocysts indicate bi-allelic gene expression during bovine preimplantation development. Gene expression activity was independent of the embryonic size.

Exons 4 and 5 and introns 4 and 5 identified within the bovine IGF2 gene are highly conserved among mammals and support the hypothesis of ROTWEIN and

HALL (1990) that sequence and structure of the IGF2 gene were developed prior to the evolutionary dispersion of mammals.

A Differentially Methylated Region (DMR) was identified in exon 10 of the bovine IGF2 gene. This final exon of the gene was hypermethylated in sperm DNA but low methylated in oocyte DNA. It is assumed that the bovine intragenic DMR corresponds to the DMR2 of the murine Igf2 gene due to high sequence conservation between different mammalian species.

The intragenic DMR2 is demethylated after fertilization on the paternally methylated chromosome as it is described in the mouse. Unlike the mouse, demethylation is completed at the 4-cell stage of bovine embryonic development.

Bovine day 7 expanded blastocysts are in an identical manner remethylated irrespective of the different origins. Thus, the in vitro production protocols had no effect on the methylation reprogramming. In contrast, hypermethylation of male cloned blastocysts is indicative for insufficient reprogramming and a sex specific methylation reprogramming is suggested. Androgenetic and parthenogenetic expanded blastocysts and adult fibroblasts showed the expected methylation levels.

The DMR is low methylated in parthenogenetic embryos due to their two maternal genomes but highly methylated in androgenetic embryos (two paternal genomes).

Somatic cells are often hypermethylated.

A Differentially Methylated Region was identified for the first time in the bovine IGF2 gene. In addition, this is the first study on analyses of methylation patterns of the bovine IGF2 gene in preimplantation embryos. Results demonstrate that the determination of methylation levels can be used as a diagnostic tool to detect aberrations in methylation reprogramming of in vitro produced embryos.

7 ZUSAMMENFASSUNG

Claudia Gebert

Untersuchungen zum Expressions- und Methylierungsmuster des bovinen IGF2 Gens in Keimzellen und präimplantatorischen Embryonen

verschiedener Herkunft

Die Methylierung der DNA an Cytosinmolekülen ist einer der am besten verstandenen Mechanismen innerhalb des epigenetischen Phänomens Imprinting.

Gene, die dem Imprinting unterliegen, spielen eine bedeutende Rolle während der embryonalen Säugerentwicklung. Die meisten Erkenntnisse über das Imprinting stammen aus Untersuchungen an Mäusen und humanen Zellen. Informationen über das Imprinting bei landwirtschaftlichen Nutztieren gibt es dagegen nur wenige. Ziel dieser Arbeit war die Untersuchung der DNA-Methylierung des Insulin-ähnlichen Wachstumsfaktor2-Gens (IGF2) in bovinen Keimzellen und präimplantatorischen Embryonen. Zusätzlich wurden Genexpressionsanalysen des IGF2, des Insulin-ähnlichen Wachstumsfaktor2 Rezeptor-Gens (IGF2R) und des Mammalian achaete-scute homolog2-Gens (MASH2) in Rinderblastozysten durchgeführt.

Die Expressionsmuster von bovinen in vitro fertilisierten (IVF) oder parthenogenetischen (PA) Blastozysten und expandierten Blastozysten an Tag 7 wurden mittels semi-quantitativer RT-PCR ermittelt. Die In-vitro-Kultur der Embryonen erfolgte im halbdefinierten SOF/BSA-Medium (Synthetic Oviduct Fluid/Bovine Serum Albumin). Über die Ermittlung des relativen Gehalts an Transkriptionsprodukten kann indirekt auf eine tendentiell mono- oder biallelische Genexpression geschlossen werden. Paternal exprimierte Gene würden dabei einen erhöhten relativen mRNA Gehalt in in vitro fertilisierten (mütterliches und väterliches

Die Expressionsmuster von bovinen in vitro fertilisierten (IVF) oder parthenogenetischen (PA) Blastozysten und expandierten Blastozysten an Tag 7 wurden mittels semi-quantitativer RT-PCR ermittelt. Die In-vitro-Kultur der Embryonen erfolgte im halbdefinierten SOF/BSA-Medium (Synthetic Oviduct Fluid/Bovine Serum Albumin). Über die Ermittlung des relativen Gehalts an Transkriptionsprodukten kann indirekt auf eine tendentiell mono- oder biallelische Genexpression geschlossen werden. Paternal exprimierte Gene würden dabei einen erhöhten relativen mRNA Gehalt in in vitro fertilisierten (mütterliches und väterliches

Im Dokument Characterization of gene expression and methylation patterns of the bovine IGF2 gene in gametes and preimplantation embryos of different origins (Seite 128-195)